Skip to the content.

Are you into volatility watching? Are you looking for the short end of the curve … shorter than one day? You found it! Peek at yarx streams and perhaps xray too. An approximate description is the code that generates them but TBH that isn’t quite the code that runs.

Oh yeah. There are prizes on offer for aggregate performance as you may have noticed on the sub-leaderboards.

Stock and protfolio streams

These are intended to provide “xray vision” into the joint behaviour of stocks. The portfolio weights are provided as JSON.

Type Example stream Lookup Reverse lookup
Stocks stream=quick_yarx_googl xraytickers.json xraytickersreverse.json
Portfolios stream=quick_xray_3 portfolios.json  
Stocks stream=middling_yarx_googl xraytickers.json xraytickersreverse.json
Portfolios stream=middling_xray_3 portfolios.json  
Stocks stream=slow_yarx_googl xraytickers.json xraytickersreverse.json
Portfolios stream=slow_xray_3 portfolios.json  

Only 1 hour ahead predictions are considered relevant for prizes.

Volatility streams

There are derived feature streams

Type Example stream Meaning
Vol stream=yarx_vlty_0_aapl Scaled log of middling_yarx_appl predictions interquartile range
Vol slope stream=yarx_vlty_1_aapl Difference between scaled log quick and log slow inter-quartile ranges
Vol convexity stream=yarx_vlty_2_aapl Similar, but the convexity

Exogenous data streams

Some exogenous features are being added that tick infrequently. They may be useful in your models, however.

Type Example stream Meaning
Vol stream=yarx_exog_0_aapl Vol feature

(ignore any exog_yarx streams they are deprecated)


Roughly, the top 100 stocks.

from import get_xray_tickers
from import get_xray_portfolios
symbols = get_xray_tickers()
weights = get_xray_portfolios()

View source